ruote log_2010-03-16

2010-03-16 05:45:27 utc jmettraux kennethkalmer: hi !
2010-03-16 05:45:43 utc jmettraux your patch is in
2010-03-16 05:45:56 utc kennethkalmer jmettraux: morning and thanks !
2010-03-16 05:45:57 utc jmettraux I changed __dispatch_time__ to dispatched_at though
2010-03-16 05:46:02 utc jmettraux thanks to you !
2010-03-16 05:46:25 utc kennethkalmer np ;)
2010-03-16 05:46:40 utc kennethkalmer i'll be wrapping up /errors at the office this morning
2010-03-16 05:47:48 utc jmettraux great
2010-03-16 05:47:59 utc jmettraux replay ftw
2010-03-16 05:48:51 utc kennethkalmer http://gist.github.com/333679
2010-03-16 05:48:59 utc kennethkalmer the pending specs
2010-03-16 05:49:05 utc kennethkalmer i have to leave for the office
2010-03-16 05:49:07 utc kennethkalmer ttyl @
2010-03-16 05:49:12 utc jmettraux :)
2010-03-16 06:44:52 utc jmettraux kennethkalmer: before I forget and it becomes a pain to change : replay is a DELETE
2010-03-16 06:45:05 utc jmettraux it really removes the current error from the log and replays at that point
2010-03-16 06:45:24 utc kennethkalmer i thought about it, you're right
2010-03-16 06:45:42 utc kennethkalmer and updating the expression needs to happen through /expressions
2010-03-16 06:45:48 utc kennethkalmer (update before replay)
2010-03-16 06:47:26 utc jmettraux +1
2010-03-16 06:48:53 utc jmettraux ruote 2.1.x has a FlowExpressionId#to_storage_id method which is URI safe
2010-03-16 06:49:52 utc kennethkalmer will dive in and check it out
2010-03-16 06:50:17 utc jmettraux so you could have /expressions/#{fei.to_storage_id}
2010-03-16 06:50:39 utc jmettraux but if you prefer wfid/expid (and it's consistent) it's fine
2010-03-16 07:14:26 utc jmettraux speaking of /expressions/wfid/expid there is also the Engine#reply(exp) thing
2010-03-16 07:17:28 utc jmettraux POST maybe
2010-03-16 07:25:40 utc kennethkalmer i'll get to /expressions later, that is going to be another can of worms to open ;)
2010-03-16 07:25:50 utc jmettraux :)
2010-03-16 09:25:59 utc der-dokctor Hi there! Anyone alive? :-)
2010-03-16 09:27:46 utc tosch_le Hi der-dokctor
2010-03-16 09:27:54 utc tosch_le yes, indeed ;-)
2010-03-16 09:31:53 utc der-dokctor Hi, tosch_le !
2010-03-16 09:32:13 utc der-dokctor good, I have a (hopfully) small question.
2010-03-16 09:32:58 utc der-dokctor in the storageParticipant's cancel Method – wht is the 'flavour' argument for?
2010-03-16 09:34:13 utc tosch_le let me have a look
2010-03-16 09:35:50 utc der-dokctor It's here http://ruote.rubyforge.org/ruote_rdoc/Ruote/StorageParticipant.html#cancel-instance_method
2010-03-16 09:35:58 utc tosch_le ay, that's a good question
2010-03-16 09:36:05 utc der-dokctor :-)
2010-03-16 09:36:20 utc tosch_le john will now better, so it will be best for you to post a question on the ml
2010-03-16 09:36:27 utc tosch_le s/now/know
2010-03-16 09:38:17 utc der-dokctor okay, thanks. Du you know how to use the method? Simply inserting '' as a value doesnt work
2010-03-16 09:39:15 utc tosch_le what's the error message/are the symptoms?
2010-03-16 09:39:58 utc der-dokctor undefined method 'to_h' for …
2010-03-16 09:40:19 utc der-dokctor in line 80
2010-03-16 09:40:46 utc tosch_le so it's complaining about the first param, not the last
2010-03-16 09:41:30 utc tosch_le the first param has to be an instance of Ruote::FlowExpressionId
2010-03-16 09:41:42 utc der-dokctor oh, it's a hash
2010-03-16 09:41:49 utc tosch_le (or deliver sensible results for calling to_h)
2010-03-16 09:41:50 utc der-dokctor in my code
2010-03-16 09:42:59 utc tosch_le just send it through Ruote::FlowExpressionId.new(your_hash)
2010-03-16 09:43:36 utc tosch_le but perhaps john should about that method to accept hashes, too. ==> ml is your friend
2010-03-16 09:44:31 utc der-dokctor oops, I'm late – I try it out later. See you! Dankeschön!
2010-03-16 12:21:22 utc tosch_le jmettraux: could you have a look at the irc log? der-dokctor wanted to know what the 'flavour' argument in Ruote::StorageParticipant#cancel is for -- and i want to know now, too ;-)
2010-03-16 12:21:53 utc jmettraux hello, unfortunately I have no access to the log from here
2010-03-16 12:22:07 utc jmettraux but I can reply
2010-03-16 12:26:29 utc jmettraux basically, for a cancel, there are 3 flavours : nil, 'kill' and 'timeout'
2010-03-16 12:26:35 utc jmettraux nil --> vanilla cancel
2010-03-16 12:26:53 utc jmettraux kill --> cancel but do not trigger any on_cancel
2010-03-16 12:27:01 utc jmettraux timeout --> the cause of the cancel was a timeout
2010-03-16 12:27:42 utc jmettraux that's it
2010-03-16 12:45:37 utc tosch_le ah, ok
2010-03-16 12:49:57 utc jmettraux it's rather "fyi", but some participant implementations could react differently based on that info
2010-03-16 12:51:00 utc tosch_le the implementation of #cancel in StorageParticipant seems to ignore the flavour argument at all
2010-03-16 12:51:09 utc jmettraux indeed
2010-03-16 12:51:25 utc tosch_le der-dokctor had also troubles using a hash as fei
2010-03-16 12:51:53 utc jmettraux I wonder what he was trying to do
2010-03-16 12:52:44 utc tosch_le he wanted to use cancel on the storageparticipant. don't ask me why. i told him to post his question on the ml, so there's a chance to know it in the end...
2010-03-16 12:52:56 utc jmettraux :) many thanks !
2010-03-16 12:58:06 utc jmettraux der-docktor: hi
2010-03-16 12:58:17 utc jmettraux 21:29 jmettraux: basically, for a cancel, there are 3 flavours : nil, 'kill' and 'timeout'
2010-03-16 12:58:18 utc jmettraux 21:29 jmettraux: nil --> vanilla cancel
2010-03-16 12:58:19 utc jmettraux 21:29 jmettraux: kill --> cancel but do not trigger any on_cancel
2010-03-16 12:58:21 utc jmettraux 21:30 jmettraux: timeout --> the cause of the cancel was a timeout
2010-03-16 13:03:02 utc der-dokctor hi john!
2010-03-16 13:03:08 utc der-dokctor thank you!
2010-03-16 13:03:30 utc der-dokctor what does the second mean?
2010-03-16 13:03:57 utc der-dokctor "not trigger"?
2010-03-16 13:04:19 utc jmettraux it means that the process was killed, not just cancelled
2010-03-16 13:04:29 utc jmettraux any :on_cancel will be ignored
2010-03-16 13:04:30 utc jmettraux http://ruote.rubyforge.org/common_attributes.html#on_cancel
2010-03-16 13:05:48 utc jmettraux 21:52 jmettraux: it's rather "fyi", but some participant implementations could react differently based on that info
2010-03-16 13:05:48 utc jmettraux 21:54 tosch_le: the implementation of #cancel in StorageParticipant seems to ignore the flavour argument at all
2010-03-16 13:05:52 utc jmettraux 21:54 jmettraux: indeed
2010-03-16 13:06:39 utc tosch_le der-dokctor: why do you want to call StorageParticipant#cancel for yourself?
2010-03-16 13:06:54 utc tosch_le s/for/by
2010-03-16 13:08:29 utc der-dokctor tosch_le: I tried to install ruote-kit into vedor, but that didn't work (I cannot get _ruote anymore and kill them with ruote-kit). Now I have plenty of workitems flying around and want to delet them ;-)
2010-03-16 13:09:15 utc jmettraux calling "cancel" isn't meant for those cases
2010-03-16 13:09:30 utc jmettraux it's called by the engine when a process is cancelled
2010-03-16 13:09:31 utc der-dokctor brb
2010-03-16 13:09:31 utc tosch_le if you want a fresh start, delete the entire work_* directory structure.
2010-03-16 13:09:37 utc jmettraux +1
2010-03-16 13:10:39 utc tosch_le if you want to clean up only some processes, use the cancel on the process (or expression) instances
2010-03-16 13:10:48 utc jmettraux +1
2010-03-16 13:11:44 utc jmettraux cancelling a workitem is an operation performed by the engine